Cloud Cost Analyses and ‘Inherent Conflict of Interest’
March 29, 2011
IT Business Edge blogger Art Cole recently wrote that "the simple economics behind cloud computing is enough to foretell its imminent rise to data center dominance," citing data from Quest Software’s Dmitry Sotnikov that shows many popular cloud services are bargains based solely on the energy needed to run a typical in-house server — never mind cooling, maintenance, security and backup costs.
The actual point of Art’s post is that moving to the cloud involves considerations beyond cost. True enough, but many IT organizations have yet to move beyond those initial cost calculations.
And quite a few of them find cloud economics far from simple. As I wrote back in July, many folks are still challenged by the idea of determining cloud costs, a confusion not helped by the growing array of cloud services and possible use cases for the cloud. Many organizations struggle with the idea of shifting IT expense from capital to operational spending.
The best way to determine cloud costs is to enlist the services of an objective third party, advises Ray DePena on the Cloud Computing Journal.
DePena, an IBM veteran and well-regarded cloud expert with one of the best LinkedIn bios ever (his current title is American Cloudsmith and Patriot in the Cloud Revolution), says it’s unrealistic to expect IT organizations to provide an unbiased comparison between on-premise and cloud costs. There’s an "inherent conflict of interest when a CEO/CFO asks the CIO to show them how they can migrate significant portions of the IT business unit to an alternative provider."
For the same reason, the third party shouldn’t be a vendor, selling either legacy infrastructure or cloud services. A vendor will hardly be "objective," as it has a vested interest in getting IT organizations to open their checkbooks.
A third party hopefully will include costs too often omitted in internal analyses, writes DePena, noting he’s seen comparative analyses with no mention of labor costs. Some internal analyses simply compare an on-premise virtual machine to a VM in the cloud and tweak the numbers by creating VM sprawl. DePena writes:
Sure, with a narrow analysis, if you take your $100,000+ system and divide by hundreds or thousands of underutilized VMs, your IT shop may artificially look competitive, though I doubt your CFO will be fooled by such a comparison.
DePena shares perhaps his most important point near the end of his post: If IT organizations want to optimize cloud savings and benefits, they’ll have to change their processes. He writes:
It doesn’t do much good to be able to bring up a VM in seconds or minutes if you have a 4-12 week legacy procurement process that no longer involves acquiring physical hardware, but VMs, while still requiring numerous management and finance approvals for what may be a temporary VM instance (and amount to a small expenditure). In my experience, the process at many enterprises take the same length of time whether it’s a 5 or 50 server purchase. Those processes will have to be re-engineered if one is serious about conducting business in the cloud.
I got a very similar take from Bernard Golden, CEO of consulting company HyperStratus and author of “Virtualization for Dummies,” when I interviewed him for a story on cloud computing back in 2009.
He told me some IT organizations might chafe at process changes required for them to enjoy the benefits of cloud computing. The cloud will call for shifts in processes as basic as budgeting, which will no longer involve submitting funding requests based upon relatively simple hardware and software cost calculations. Said Golden:
If IT organizations want to take advantage of cloud computing, they will have to modify or even jettison many of their existing processes.